Lithiated anions derived from (alkenyl)pentamethyl phosphoric triamides: An accurate study of the carbanion formation mechanism

Article abstract of DOI:10.1016/j.jorganchem.2004.02.003

The mechanism of the formation of lithiated carbanions derived from (alkenyl) pentamethyl phosphoric triamides has been elucidated. Whereas a few initial ambident allylphosphoramide anion was formed with n-BuLi, both reversible α-reprotonation and not reversible γ-reprotonation simultaneously occurred as a result of the reaction between the ambident carbanion formed and the starting enephosphoramide. A such autocatalytic process led partially to the transposed allylphosphoramide isomer. In the case of α-phenyl substituent the transposed phosphoramide was not a difficulty because it was further γ-deprotonated in situ with the n-BuLi still present, provided finally the expected ambident anion. With α-methyl and α-propyl substituent the transposed enephosphoramide formed in the autocatalytic process was not γ-deprotonated and consequently was prejudicial to the preparation of the target ambident carbanion. In these last cases, adapted experimental conditions avoided the autocatalytic process and allowed the preparation of the corresponding anions.

Full text of DOI:10.1016/j.jorganchem.2004.02.003

Journal of Organometallic Chemistry 689 (2004) 1530–1539
www.elsevier.com/locate/jorganchem
Lithiated anions derived from (alkenyl)pentamethyl phosphoric
triamides: an accurate study of the carbanion formation mechanism
1
Claude Grison, Antoine Thomas , Frederic Coutrot, Philippe Coutrot
*
ꢀ ꢀ
Laboratoire de Chimie Organique Biomolꢀeculaire, Institut Nancꢀeien de Chimie Molꢀeculaire, FR CNRS 1742, UMR 7565,
Universitꢀe Henri Poincarꢀe, Nancy 1, BP 239, 54506 Vandoeuvre cꢀedex, France
Received 15 December 2003; accepted 3 February 2004
Abstract
The mechanism of the formation of lithiated carbanions derived from (alkenyl) pentamethyl phosphoric triamides has been
elucidated. Whereas a few initial ambident allylphosphoramide anion was formed with n-BuLi, both reversible a-reprotonation and
not reversible c-reprotonation simultaneously occurred as a result of the reaction between the ambident carbanion formed and the
starting enephosphoramide. A such autocatalytic process led partially to the transposed allylphosphoramide isomer. In the case of
a-phenyl substituent the transposed phosphoramide was not a difficulty because it was further c-deprotonated in situ with the n-
BuLi still present, provided finally the expected ambident anion. With a-methyl and a-propyl substituent the transposed ene-
phosphoramide formed in the autocatalytic process was not c-deprotonated and consequently was prejudicial to the preparation of
the target ambident carbanion. In these last cases, adapted experimental conditions avoided the autocatalytic process and allowed
the preparation of the corresponding anions.
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Enephosphoramide; Allylic transposition; Lithium reagents; Alkenyl pentamethyl phosphoric triamides; Ambident anions
1. Introduction
starting material, reduced stability of the allyl anions
and not easy hydrolysis of the c-adduct [3].
The heteroatom-assisted c-deprotonation of allylic
group is a convenient and efficient strategy for the
preparation of aldehyde homoenolate synthons [1].
Heterosubstituent is used to stabilize the allylic anion
and to generate a masked carbonyl group after c-reac-
tion with electrophilic reagents. Heterosubstituted allyl
anions belong now to the synthetically important group
of reversed polarity synthons and are versatile tools for
further synthetic transformations [2].
As part of a program aimed at developing new ho-
moenolate synthons we have previously reported the
synthetic utility of lithium anions derived from allyl-
phosphoramides as aldehyde [4] and recently ketone
homoenolate equivalents [5]. In this work, we describe
accurately the different interesting facts observed during
the formation of the ketone homoenolate equivalents
derived from N-[(1-alkyl or 1-aryl)-2-propenyl] phos-
phoramides.
The use of allylic anions as ketone homoenolate
equivalents however is far less documented. It brings
new problems: difficult or impossible access to the
2. Results and discussion
1
2.1. Stability of lithiated carbanions 3 and H and 31P
NMR data of starting enephosphoramides 2, carbanions 3
and transposed enephosphoramides 4 (Z/E)
* Corresponding author. Tel.: +33-3-83-68-43-61; fax: +33-3-83-68-
43-63.
E-mail address: philippe.coutrot@incm.uhp-nancy.fr (P. Coutrot).
1
The starting compounds (alkenyl) pentamethyl
phosphoric triamides 2 and the lithiated anions derived
Part of the thesis of author; Thomas, A. Thesis, University Henri
ꢀ
Poincare, Nancy 1, France, 2002.
0022-328X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.02.003

C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
1531
Me
N
H
N
iii) n-BuLi/THF
iv) MeI
H
N
ii) Me2NH, Et3N
(Me2N)2P
i) POCl3, Et3N
(Me2N)2P
H2N
Cl2P
O
O
O
R
R
R
R
2
1
n-BuLi
THF/hexane/-50˚C
Me
Me
N
H2O
α
γ
(Me2N)2P
O
N
, Li
(Me2N)2P
O
R
R
4
3
Scheme 1.
Table 2
Stability of the anions 3
3 were prepared according to Scheme 1 as previously
described [5].
31P RMN established the quantitative formation of
anions 3: the signal of anions 3 was ꢀ1 ppm downfield
from that of the phosphoramides 2 in THF (Table 1).
The stability of the anions 3 was very temperature-
dependent. The metallation of 2c could be achieved
smoothly at )10 °C leading to the anion 3c which was
perfectly stable at 20 °C and kept at 20 °C during 4 h
without degradation. As expected, the anions 2a and 2b
were less stable and it was obliged to keep the reaction
medium at low temperature (Table 2).
Entry
3
R
Temperature (°C) Time
1
2
3a
3a
3a
3b
3b
3b
3b
3c
3c
3c
Me
Me
Me
Pr
)50
)10
+20
)50
)20
0
>2 h
A few minutes
3
Immediate degradation
>2 h
4
5
Pr
Pr
2 h
1 h
6
7
Pr
+20
)50
+20
+58
Immediate degradation
8
Ph
Ph
Ph
>2 h
>4 h
1 h
9
10
Hydrolysis or deuterolysis at one sweep at )50 °C of
carbanions 3 gave exclusively the fully transposed ene-
phosphoramides 4 contrary to lithiated a-unsubstituted
enephosphoramides (R ¼ H) which were known to lead
after hydrolysis to a mixture of transposed and not
transposed enephosphoramide products. The double
bond configuration of transposed enephosphoramides
4a-b depended on the hydrolysis temperature. Rapid
hydrolysis at )50 °C gave the (Z) enephosphoramides
4a-b which are the kinetic products (Scheme 2). Those
results could be explained by an internal lithium che-
lating stabilising effect of the phosphoramide group into
3a-b. Temperature increase or addition of HMPA pre-
vented the stabilisation and led to a mixture of (Z) and
(E) products. In the case of 3c the sole transposed (Z) 4c
was obtained whatever the hydrolysis temperature, the
phenyl hindrance forced the (Z) 3c configuration which
was maintained after hydrolysis.
1H and 31P NMR spectra allowed a clear identifica-
tion of transposed 4 and not transposed 2 enephos-
phoramides. In the 31P NMR spectra (in CDCl3) a
singlet was observed at 22.68–22.80 ppm for the starting
enephosphoramides 2 whereas in the transposed product
4, this signal appeared in all cases upfield at 18.00–19.57
(Table 1, Fig. 1). Moreover the distinction between 2
and 4 was facilitated by the observation of the Ca–N–
Me chemical shift which constituted an efficient internal
probe. In the 1H NMR spectra the Ca–N–Me signal
appeared as a doublet at 2.42–2.46 ppm for 2
(3JH–P ¼ 8:9–9.5 Hz) whereas in the transposed product
Table 1
31P NMR chemical shifts of 2, 3 and 4
d (ppm) (THF)
Entry
R
2
d (ppm)
(CDCl3)
22.7
3
4
d (ppm)
(THF)
21.0
(CDCl3)
(THF)
1
2
3
Me
Pr
2a
2b
2c
3a
3b
3c
21.9
22.3
22.7
(Z) 4a
(E) 4a
(Z) 4b
(E) 4b
(Z) 4c
(E) 4c
18.19
19.40
18.13
19.57
18.00
22.7
22.8
21.0
20.9
Ph
16.5
17.4

1532
C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
Me
N
Li
Me
Me
(Me2N)2P
N
N
(Me2N)2P
O
n-BuLi
(Me2N)2P
O
O
Li
R
R
R
2
(E) 3
(Z) 3
H2O
E
Me
(E) 3a-b
(Z) 3a-b
N
(Me2N)2P
O
4
R
T˚C
Z / E
96 / 4
T˚C
Z / E
4a
4b
4c
-50˚C
- 10˚C
“
40 / 60
45 / 55
100 : 0
“
“
100 / 0
100 / 0
+ 20˚C
2a-b
(Z) 4a-b
(E) 4a-b
reaction coordinate
Scheme 2.
4 it was observed in all cases at 2.72–2.94 ppm
(3JH–P ¼ 8:0–8.7 Hz) (Fig. 2).
product) the splitting of this signal was characteristic of
the alkylation product. If it presented as a quadruplet it
was the transposed enephosphoramide resulting of the
hydrolysis of the not reacted carbanion 3a. If this signal
appeared as a triplet it was the alkylation product 5a.
The configuration of the double bond into the
transposed enephosphoramides 4 was also assigned
from the 1H and 31P NMR spectra as exemplified in the
case of (Z) 4a and (E) 4a as well as in the case of (Z) 5a
obtained by methylation of the carbanion 3a with
iodomethane (Fig. 1 and 2).
2.2. More accurate 31P NMR study of the reaction
mechanism between n-BuLi and enephosphoramides 2a-c
for the obtention of carbanions [3a-c] at )50 °C
In 31P NMR (in CDCl3) the signal of (Z) 4a was lo-
cated at 18.19 ppm whereas that of (E) 4a was slightly
deshielded at 19.4 ppm (Table 1, Fig. 1). The observa-
2.2.1. Reaction between n-BuLi and enephosphoramide 2c
Addition of n-BuLi to the enephosphoramide 2c
at )50 °C in THF led to the immediate and total for-
mation of carbanion [3c] as it could be observed from
the presence of the sole signal at 22.7 ppm in 31P NMR
just after the total introduction of one equivalent of n-
BuLi (Table 2, Fig. 3, D). However the monitoring of
the reaction according to the amount of n-BuLi added,
shown two phases.
During the first half of the addition, it was observed a
continue chemical shift variation from the initial signal
at 20.9 ppm characteristic of 2c towards that of [3c] at
22.7 ppm, indicating a coalescence between the two
signals. In the same time, two new signals appeared at
17.40 and 16.50 ppm, respectively, assigned to trans-
posed enephosphoramides (E) 4c and (Z) 4c. Until half-
addition these signals progressively increased at these
invariable chemical shifts (Fig. 3, A–C). At half-addition
4
1
tion of the J allylic coupling in H NMR of the two
split signals (d) and (f) is also characteristic of (Z) or (E)
4a (Fig. 2). The signal (d) of pure (Z) 4a appeared as a
quadruplet mainly resulting from a 3J coupling with the
4
Me (e). The cis J allylic coupling between (d) and (f)
was poor and did not appear. On the contrary, the same
signal (d) into (E) 4a presented a multiplet correspond-
3
4
ing to both coupling JH–Me ðeÞ and JH–Me ðfÞ. This ob-
servation was confirmed by new splitting of signals (f)
and (e).
It was also noteworthy that the both appearance of
the signal (d) and the disappearance of the signal of the
hydrogen (c) are of great utility to estimate the conver-
sion into the transposed alkylation product as 5a. Al-
though the chemical shifts of (d) were very close into 4a
(product of hydrolysis of the carbanion 3a having not
reacted with the alkylating reagent) and 5a (alkylation

C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
1533
2a
(Z) 4a
(E) 4a
(Z) 4a
(Z) 5a
34
32
30
28
26
24
22
20
18
16
14
12
(ppm)
Fig. 1. 31P NMR spectra (CDCl3) of enephosphoramides 2a (starting compound), 4a (transposed product), 5a (methylation product).
of n-BuLi the reaction medium contained approximately
50% of carbanion [3c] and 50% of enephosphoramide
4c. The addition going on, the signals of 4c progressively
disappeared and at the end of addition have completely
disappeared (Fig. 3, D).
phosphorus atom. Appearance of transposed enephos-
phoramide (E, Z) 4c can be explained by a partial ir-
reversible c-reprotonation of [3c] with 2c. An other
experience consisting in the addition of a small amount
of 2c to the sole carbanion [3c] proves such a suppo-
sition. In that attempt it was effectively observed an
upfield shift variation of the signal of [3c] from 22.7 to
22.4 ppm and the signal of (E) 4c at 17.40 ppm as
expected. It has to note that the formation of major
(E) 4c in the reaction medium during the first part of
addition of n-BuLi seems contrary to the result of
hydrolysis of the carbanion [3c] which provides (Z) 4c
as related above. Consequently we suppose that the
starting compound 2c can induce a destabilising effect,
These different observations could be explained by
the following mechanism (Fig. 4).
During the first period the ambident carbanion [3c]
behaved as a transposition catalyst. The coalescence of
signals of 2c and [3c] shows that an acid–base equi-
librium occurs between 2c and [3c] and that the a-
reprotonation of [3c] with 2c is entirely reversible. It
indicates also that the hydrogen exchange rate between
both species is faster than the relaxation time of

1534
C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
(a)
(b)
(d)
H
(f)
Me
(b)
(a)
(Me2N)2P
(c)
(e)
N
H
(e, e')
O
(d)
Me
(c)
H
(f)
(e')
TMS
2a
(a)
(a)
(a)
(b)
(e)
(f)
Me
Me
(a)
(b)
N
(Me2N)2P
H
(d)
(e)
O
(d)
Me
(f)
TMS
(Z) 4a
(f)
(b)
Me
(d)
H
(a)
N
(Me2N)2P
(b)
(e)
(e)
(d)
Me
O
Me
TMS
(f)
(Z + E) 4a
(g)
(b)
(e)
CH3
(f)
(g)
Me
N
(a)
(Me2N)2P
(b)
H
(d)
TMS
(d)
(e)
O
Me
(f)
(Z) 5a
6.0
5.0
4.0
3.0
(ppm)
2.0
1.0
0.0
Fig. 2. 1H NMR spectra (CDCl3) of 2a (starting compound), (Z) (E) 4a (transposed products), (Z) 5a (methylation product).
like HMPA, on the intermediate (Z) [3c] which would
lead to (E) 4c, in situ.
establish that neither reversible reaction exist between
both species. A smallest proton rate exchange between
[3c] and 4c compared to the relaxation time of phos-
phorus atom could also explained this fact. The coa-
lescence signal which reached its maximum at 22.7 ppm
at half-addition and which maintained this value until
the end of addition was the best argument. A reproto-
nation of [3c] by 4c after the half-addition would pro-
duce again 2c in the reaction medium and consequently
a new coalescence between the signals of 2c and [3c]
would exist involving an upfield variation of that signal,
that was not the case.
Finally it has to be noted that this study that revealed
the presence of different intermediates during the met-
allation of 2c can induce the idea of a long time metal-
lation. It is not the case. With a synthetic aim, carbanion
[3c] was fully formed at )50 °C in THF after addition in
After the half-equivalence, the progressive disap-
pearance of signals 4c was observed without any change
in the chemical shift of [3c] which remained the unique
signal at the end of addition. At this time, one equiva-
lent of n-BuLi was added and 100% of carbanion [3c]
was sole present in the medium.
That means that the addition of n-BuLi from half-
reaction only used to deprotonate the transposed
phosphoramide 4c into the ambident anion [3c]. This
was verified in a parallel experience consisting to test the
ability for n-BuLi to effect the c-deprotonation of pure
4c followed by a subsequent deuterolysis that afforded c-
deuterated transposed enephosphoramide 4c.
The absence of coalescence between the signals of [3c]
and 4c cannot sole constituted a sufficient prove to

C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
1535
2c
A : 0% de n-BuLi
2c + [3c]
(E) 4c
B : 25% de n-BuLi
[3c]
[3c]
(E) 4c
(Z) 4c
C : 50% de n-BuLi
D : 100% de n-BuLi
24
22
20
18
16
(ppm)
Fig. 3. n-BuLi addition-evolution of different species observed in the reaction with enephosphoramide 2c at )50 °C (31P NMR in THF in mode
sweep-off).
Me
N
(Me2N)2P
O
n-BuLi
+
Ph
2c
-50˚C / THF
Me
Me
N
(Me2N)2P
N
(Me2N)2P
O
γ
α
+
n-BuH
+
until
O
half-introduction
of n-BuLi
Li
Ph
Ph
[3c]
2c
γ-reprotonation
α-reprotonation
(not reversible)
(reversible)
Me
(Me2N)2P
O
N
2c
+ [3c]
+
[3c]
Ph
H
4c
n-Bu Li
after
half-introduction
of n-BuLi
Me
N
(Me2N)2P
O
Li
Ph
[3c]
Fig. 4. Mechanism of the reaction of enephosphoramide 2c with n-BuLi.
1 min. of 2 ml of n-BuLi (2.5 M in hexane, 5 mmol, 1.1
equiv.) to 4.5 mmol of enephosphoramide 2c in 35 ml of
THF. Immediate deuterolysis of the reaction mixture or
methylation with iodomethane respectively yielded
100% of transposed c-deuterated (Z) 4c or 100% of (Z)
5c.

1536
C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
2.2.2. Reaction between n-BuLi and enephosphoramide 2a
The metallation of 2a with n-BuLi was as swift as 2c
in the same precedent conditions since after addition of
n-BuLi in 1 min to 2a and subsequent immediate deu-
terolysis, 100% of transposed c-deuterated (Z) 4a was
obtained.
2.3. Reaction between n-BuLi and enephosphoramide 2b
2.3.1. Metallation at )50 °C (Fig. 5)
Whereas the metallation of enephosphoramides 2a
and 2c was extremely rapid at )50 °C, complete de-
protonation of 2b needed 1 h at this temperature after
the complete addition of n-BuLi. 31P NMR monitoring
sweep-off of the metallation at )50 °C shown that coa-
lescence did not exist between the signals of enephos-
phoramide 2b at 21.0 ppm and that of corresponding
carbanion [3b] at 22.6 ppm whatever the amount of n-
BuLi added. Either the reversible a-deprotonation of 2b
with [3b] was not possible or the rate of the exchange of
a-hydrogen between these species was slower than the
relaxation time of phosphorus atom in 31P NMR.
Otherwise, the in situ c-reprotonation of [3b] into 4b
was never observed at )50 °C in contrast with the
metallation of 2a or 2c: in this case the transposed
enephosphoramide 4b was not present in situ at half-
addition of n-BuLi. The sole noted change in the 31P
NMR spectra of the reaction medium relatively to the n-
BuLi added was the gradual appearance of the carban-
ion signal [3b] and the concomitant, then complete,
disappearance of the signal 2b.
When n-BuLi was added slowly at )50 °C, 31P
NMR monitoring sweep-off of the metallation shown
that carbanion [3a] behaved as a transposition catalyst
like [3c]. For instance, hydrolysis of the reaction mix-
ture at half-addition afforded 100% of transposed
enephosphoramide 4a. That means that 50% of 4a was
the result of the autocatalytic process. But in contrast
with the precedent case of 4c, n-BuLi was unable to
deprotonate in situ the transposed enephosphoramide
4a. There is here an essential difference in the forma-
tion of [3a] compared to that of [3c]: in the case of [3c]
the formation of transposed product 4c which accom-
panied [3c] at half-addition was not a problem since
the enephosphoramide 4c was entirely retransformed
into [3c] in the second part of the addition of n-BuLi,
and this whatever the addition rate of n-BuLi. As a
result, 100% of added n-BuLi provided 100% of [3c].
In contrast, a slow metallation of 2a produced ene-
phosphoramide 4a at the expense of [3a] that therefore
could not be obtained in 100% yield. Consequently a
very fast addition of n-BuLi is recommended in this
case, so that a-deprotonation of 2a provides [3a] faster
than autocatalysis leading to transposed enephospho-
ramide 4a.
2.3.2. Metallation at )50 °C/1 h, then gradual warming of
the reaction mixture to 0 °C
In this experience 0.25 equiv. of n-BuLi was added
to 1 equiv. of 2b at )50 °C. After 1 h at this temper-
ature, a sample was taken and analysed by 31P NMR
2b
Spectrum A : 0% de n-BuLi
Spectrum B : 25% de n-BuLi
Spectrum C : 50% de n-BuLi
2b
[3b]
[3b]
2b
[3b]
Spectrum D : 100% de n-BuLi
40
35
30
25
(ppm)
20
15
Fig. 5. n-BuLi addition-evolution of different species observed in the reaction with enephosphoramide 2b at )50 °C (31P NMR in THF in mode
sweep-off).

C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
1537
2b
4b
[3b]
(Z)
Spectrum A (t = 0 min.)
(E)
(E)
4b
(Z)
(Z)
2b
[3b]
Spectrum B (t = 2min.)
Spectrum C (t = 60 min.)
4b
4b
(E)
(E)
(Z)
15
Spectrum D (t =180min.)
40
35
30
25
(ppm)
20
Fig. 6. Evolution of different species observed in the reaction between enephosphoramide 2b (1 equiv.) and n-BuLi (0.25 equiv.) between )50 and 0 °C
31P NMR in THF in mode sweep-off).
(
(without hydrolysis). The presence of 25% of [3b], 75%
starting compound 2b and traces of 4b was observed.
The reaction mixture was allowed to gradually warm
to 0 °C with successive analyses at different tempera-
tures. Neither change occurred until )30 °C but after
1 h at )20 °C, 10% of 4b was present. Increasing of the
temperature enhanced the autocatalytic process. After
1 h at )10 °C 33% of 4b was noted, and 100% of 4b
was finally observed at 0 °C before hydrolysis. Fig. 6
illustrates this evolution at different times with a sam-
ple taken at )50 °C and placed in the not thermostat
probe of the NMR apparatus allowing it to warm to
ambient temperature.
In summary, the carbanion [3b] (R ¼ Pr) presented a
similar behaviour that [3a] (R ¼ Me) and [3c] (R ¼ Ph)
as transposition catalyst, but it distinguished from these
carbanions by a greater auto catalysis temperature that
only efficiently occurred between )10 and 0 °C.
This auto catalysis temperature was of great impor-
tance in the preparation of carbanion [3b] since once
formed 4b, n-BuLi was unable to deprotonate 4b into [3b]
(like with 4a where n-BuLi was unable to deprotonate 4a
into [3a]). A parallel attempt shown indeed that the re-
action between 4b and n-BuLi followed by subsequent
quenching with D2O afforded 100% of not deuterated 4b.
As a result it was necessary to rigorously maintain the
temperature of the metallation of 2b with n-BuLi at )50
°C for 1 h to obtain 100% of carbanion [3b].
3. Conclusion
The 31P NMR study of the reaction of metallation of a-
substituted enephosphoramides 2a-c (R ¼ Me, Pr, Ph)
with n-BuLi allows to establish the best conditions for the
preparation of the lithiated corresponding carbanions
and determine their stabilities. The formation of a-methyl
[3a] and a-phenyl [3c] substituted carbanions is immediate
at )50 °C in THF, whereas 1 h is necessary for the met-
allation of 2b (R ¼ Pr) at this temperature. The stability of
these ambident anions vary according to their structure.
The a-phenyl [3c] substituted carbanion is by far the most
stable since it can refluxed in THF for 1 h without any
degradation. The a-propyl [3b] begins to degrade after 1 h
at 0 °C whereas the a-methyl [3a] is the less stable and fast
degrades within a few minutes at )10 °C.
A fine autocatalytic reaction has been revealed during
these different metallations leading to a plausible reaction
mechanism. It was established that the formed carba-
nions [3a-c] behaves like bases towards the starting ene-
phosphoramides 2a-c as the n-BuLi addition proceeds. In
this way these ambident anions are able to repronate at
the a- or at the c-position to respectively afford again 2a-c
or transposed enephosphoramides 4a-c.
It was demonstrated that a-reprotonation is reversible
at T P ꢁ 50 °C in the case of a-substituted phenyl ene-
phosphoramide [3c] and not reversible for [3b] whatever
the temperature between )50 and 0 °C. As a result of the

1538
C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
great fragility of [3a], the monitoring of the reaction me-
dium in a not thermostat 31P NMR probe being not pos-
sible without degradation, the eventual coalescence
between the signals of 2a and [3a] could not been observed.
Consequently, the reversibility of a-deprotonation re-
mains a not resolved question in the case of 2a. On the
other hand, c-reprotonation of [3a-c] is not reversible
whatever the nature of the carbanion, and constitutes an
auto catalysis process providing, in situ, transposed ene-
phosphoramides 4a-c. This phenomena has to be abso-
lutely avoided in the cases of a-methyl and a-propyl
enephosphoramides2a-b during themetallationleadingto
[3a-b] whereas it is not a problem in the metallation of 2c.
dichloromethane (3 ꢂ 20 ml). The combined organic
layers were dried (MgSO4) and the solvent was removed
under reduced pressure to afford the c-deuterated
phosphoramide 4c. This compound was identical to c-
deuterated phosphoramide 4c obtained according to the
same experimental conditions applied to the starting
enephosphoramide 2c [5].
[(1-phenyl-1-propen-3d1-1-yl)pentamethyl phosphoric
triamide] 4c. Yield: 100%; pale yellow oil; (Z/E: 100/0); IR
(NaCl plates)/cmꢁ1: mmax ¼ 3026, 2923, 2151, 1637, 1592,
1
1490, 1454 and 1299; H NMR (250.13 MHz, CDCl3):
dH ¼ 1:92 (m, 2H, C@CH–CH2–D), 2.55 [d, 3JH–P ¼ 9:0
3
Hz, 12H, [(CH3)2N]2PO], 2.93 [d, JH–P ¼ 8:6 Hz, 3H,
CH3–N–CH(Ph)], 5.82 [t, 3JH–H ¼ 6:8 Hz, 1H,
C(Ph)@CH–CH2D], 7.19–7.51 (m, 5H, Ph); 13C NMR
1
4. Experimental
(62.896 MHz, CDCl3): dC ¼ 13:6 (t, JC–D ¼ 19:0 Hz,
C@CH–CH2D), 36.7, 36.8 [[(CH3)2N]2PO and CH3–N–
3
IR spectra were obtained using a Nicolet 210 spec-
trometer and are given in cmꢁ1. 1H NMR/13C NMR/31P
NMR spectra were recording on a Bruker AC250. Data
CH(Pr)], 122.6 (d, JC–P ¼ 5:0 Hz, C@CH–CH2D),
126.0, 126.9, 127.8, 140.4 (Ph), 143.3 [N–C(Ph)@CH–
CH2D]; 31P NMR (101.256 MHz, CDCl3): dP ¼ 17:97;
MS (EIþ) m=z calculated for C14H23DN3OP [M]þ 282.3
found 282 [[M]þ, 70%], 147 [[M–(Me2N)2PO]þ, 78%], 135
[[(Me2N)2PO]þ, 100%].
1
for H NMR spectra are reported in d units downfield
from internal Me4Si. 13C NMR spectra were referenced
to the CDCl3 peak at 77 ppm relative to Me4Si. Or-
thophosphoric acid (85%) was used as an external
standard for 31P NMR. Multiplicities are reported as s
(singlet), d (doublet), t (triplet), q (quartet), m (multi-
plet). Mass spectra (EI, 70 eV and CI) were obtained on
a Fison Trio 1000 spectrometer. Analytical chromatog-
raphy was performed on silica gel 60 F254 plates. TLC
plates were developed with spraying sulfuric acid fol-
lowed by calcination, by iodine, or by UV. Preparative
chromatographic separations were carried out on Merck
silica gel 60 (230–400 mesh). Et2O was distilled over
P2O5 and stored over Na. THF was freshly distilled
from Na/naphtalene prior to use. Benzene was distilled
4.3. 31P NMR monitoring of the reaction between n-BuLi
and enephosphoramides 2
To a stirred solution of transposed enephosphora-
mide 2 (4.5 mmol) in THF (10 ml) at )50 °C was added
in 0.5 ml equal shares a solution of n-BuLi (2.0 ml, 5
mmol, 2.5 M in hexane). After each addition of n-BuLi,
a sample was taken off at )50 °C, placed in a NMR tube
under nitrogen atmosphere, and rapidly analysed by 31
P
NMR in mode sweep-off in a not thermostat probe.
Consequently it was not possible to assign the exact
analysis temperature which was probably greater than
)50 °C. However, these conditions were compatible
with the stability of [3b-c] in the probe so that the dif-
ferent recorded spectra could be considered significant
for the mechanism determination. In the case of the
fragile carbanion [3a] samples were cut off, then hy-
drolysed before NMR analysis.
ꢁ
over Na and stored over molecular sieves (3 A). N-Bu-
tyllithium was purchased from Aldrich and was titrated
using the Watson–Eastham procedure.
4.1. Preparation of ene phosphoramides 2, 4, and 5
Typical procedures for the preparation of ene phos-
phoramides 2, 4, and 5 were described previously with
the corresponding analytical data [5].
References
4.2. Reaction between n-BuLi and transposed enephos-
phoramide 4c
[1] Reviews for homoenolate anions equivalent, see: (a) H. Ahlbrecht,
U. Beyer, Synthesis 3 (1999) 365–390;
(b) J.C. Stowel, Chem. Rev. 84 (1984) 409–435;
(c) N.H. Werstiuck, Tetrahedron 39 (1983) 205–268.
O-substitued allyl anions: (a) D.A. Evans, G.C. Andrews, B.
Buckwalter, J. Am. Chem. Soc. 96 (1974) 5560–5561;
(b) W.C. Still, T.L. MacDonald, J. Am. Chem. Soc. 96 (1974)
5561–5563;
To a stirred solution of transposed enephosphora-
mide [5] 4c (4.5 mmol) in THF (10 ml) at )50 °C was
added a solution of n-BuLi (2.0 ml, 5 mmol, 2.5 M in
hexane). After stirring under nitrogen atmosphere at
)50 °C for 5 min, the mixture was rapidly hydrolysed
with deuterium oxide (1.5 ml) at this temperature fol-
lowed by an addition of 15 ml of a NaCl aqueous sat-
urated solution. The aqueous layer was extracted with
(c) T. Mukaiyama, T. Hayashi, T. Miwa, K. Narasaka, Chem.
Lett. (1982) 1637–1640;
(d) T. Cuvigny, M. Julia, L. Jullien, C. Rolando, Tetrahedron
Lett. 28 (1987) 2587–2590;

C. Grison et al. / Journal of Organometallic Chemistry 689 (2004) 1530–1539
(e) A. Yanagisawa, K. Yasue, H. Yamamoto, Synlett. (1993) 686– Tetrahedron Lett. 1 (1976) 47–50;
1539
688;
(f) X. Teng, Y. Takayama, S. Okamoto, F. Sato, J. Am. Chem.
Soc. 121 (1999) 11916–11917;
(b) D.A. Evans, J.M. Takacs, K.M. Hurst, J. Am. Chem. Soc. 101
(1979) 371–378.
[2] For methods umpolung, see: D. Seebach, Angew. Chem., Int. Ed.
Engl. 18 (1979) 239–336;
(g) S. Okamoto, X. Teng, S. Fujii, Y. Takayama, F. Sato, J. Am.
Chem. Soc. 123 (2001) 3462–3471.
(a) H. Ahlbrecht, G. Rauchschwalbe, Synthesis (1973) 417–
420;
S-substitued allyl anions: (a) K. Oshima, H. Yamamoto, H.
Nozaki, J. Am. Chem. Soc. 95 (1973) 7926–7928;
(b) Y. Yamamoto, H. Yatagai, K. Maruyama, J. Chem. Soc.
Chem. Commun. (1979) 157–158;
(b) C.D. Ayalon, E. Ehlinger, P. Magnus, J. Chem. Soc., Chem.
Commun. (1977) 772–773;
(c) H. Ahlbrecht, A. Kramer, Chem. Ber. 129 (1996) 1161–1168;
(d) J. Enda, I. Kuwajima, J. Am. Chem. Soc. 107 (1985) 5495–
5501.
(c) L. Lambs, N.P. Singh, J.F. Biellmann, J. Org. Chem. 57 (1992)
6301–6304.
Si-substitued allyl anions: (a) P.W.K. Lau, T.H. Chan, Tetrahe-
dron Lett. 27 (1978) 2383–2386;
[3] (a) I. Kuwajima, M. Kato, J. Chem. Soc., Chem. Commun. (1979)
708–709;
(b) A. Hosomi, H. Hashimoto, H. Sakurai, J. Org. Chem. 43
(1978) 2551–2552;
(b) D. Hoppe, Angew. Chem., Int. Ed. Engl. 23 (1984) 932–948;
(c) J. Enda, I. Kuwajima, J. Am. Chem. Soc. 107 (1985) 5495–5501;
(d) H. Ahlbrecht, R. Schmidt, U. Beyer, Eur. J. Org. Chem. (1998)
1371–1377.
(c) E. Ehlinger, P. Magnus, J. Am. Chem. Soc. 102 (1980) 5004–
5011;
(d) T.H. Chan, D. Wang, Chem. Rev. 95 (1995) 1279–1292.
N-substitued allyl anions: (a) M. Julia, A. Schouteeten, M.
Baillarge, Tetrahedron Lett. 38 (1974) 3433–3434;
(b) S.F. Martin, M.T. Dupriest, Tetrahedron Lett. 45 (1977) 3925–
3928;
[4] (a) P. Coutrot, P. Savignac, Y. Leroux, C.R. Acad. Sci. Ser. C
(1974) 279–609;
(b) P. Coutrot, P. Savignac, J. Chem. Res. (S) (1977) 308;
P. Coutrot, P. Savignac, J. Chem. Res. (M) (1977) 3401–3416;
(c) P. Coutrot, J.R. Dormoy, A. Moukimou, J. Organomet.
Chem. 258 (1983) C25–C28;
(c) T. Hassel, D. Seebach, Angew. Chem. Int. Ed. Engl. 18 (1979)
399–400;
^
(d) P. Coutrot, C. Grison, C. Bomont, Phosphorus Sulfur 77
(d) H. Ahlbrecht, G. Bonnet, D. Enders, G. Zimmermann,
Tetrahedron Lett. 21 (1980) 3175–3178;
(1993) 195;
(e) P. Coutrot, C. Grison, C. Bomont, Tetrahedron Lett. 35 (1994)
^
(e) G.A. Weisenburger, P. Beak, J. Am. Chem. Soc. 118 (1996)
12218–12219;
8381–8384;
(f) P. Coutrot, C. Grison, C. Bomont, J. Organomet. Chem. 586
^
(f) M.C. Whisler, E.D. Soli, P. Beak, Tetrahedron Lett. 41 (2000)
9527–9531.
P-substitued allyl anions: (a) G. Sturtz, B. Corbel, J.P. Paugam,
(1999) 208–217.
[5] C. Grison, A. Thomas, F. Coutrot, P. Coutrot, Tetrahedron 59
(2003) 2101–2123.